Parameter calibration for the discrete element simulation of soil-preceding rice stubble and its rotary tillage trajectory

Due to the lack of reliable discrete element simulation parameters, the motion pattern and mutual force of soil-preceding rice stubble cann't be accurately calculated during rotary tillage after rice harvest in rice-oilseed rape rotation area. The EDEM software was used in this study to calibrate the physical and contact parameters of each substance within the soil-preceding rice stubble mixture by discrete element simulation. A series of physical experiments were carried out to determine the physical and mutual contact parameters of each substance within the soil-preceding rice stubble mixture, including density, modulus of elasticity, Poisson's ratio, static friction parameter, rolling friction parameter, and collision recovery coefficient. Taking the accumulation angle as the evaluation index, the Plackett-Burman experiment was used to screen the significant factors, and the range of factors was further optimized by the steepest-climbing experiment, and the Box-Behnken experiment was further carried out within the optimized range, to observe the effects of the factors and interactions on the model, and then a second-order regression model was obtained to search for the optimal parameters and to complete the model calibration. The optimal parameter combination of the most significant parameters of the soil-preceding rice stubble model were soil-soil recovery coefficient 0.407, soil-soil kinetic friction coefficient 0.123, soil-preceding rice stubble static friction coefficient 0.596, and the soil surface energy 1.860 J, The error of actual accumulation angle was 0.58% under the optimal parameter combination. Direct shear experiments and simulations for the soil-preceding rice stubble were carried out, in which a scenario of preceding rice stubble crumb embedded into the soil consistent with the physical experiment was observed in the simulation. The shear stress errors of the simulated and experimented direct shear stresses at different vertical loadings after vertical stress stabilization were obtained as 5.4% (50 kPa), 4.1% (100 kPa), and 3.1% (150 kPa), respectively. The reasons for the errors under different vertical stresses were analyzed from both the model itself and the experiments, and the accuracy of the simulation model's contact stresses was verified. Under the optimal parameter combination, the MBD-DEM-based rotary soil-preceding rice stubble trajectory analysis experiment was carried out, and two soil-preceding rice stubble burial scenarios were summarized, in which it was observed that the rice preceding stubble would fall in the opposite direction to the forward direction of the rotary tillage and be squeezed by the subsequent knives to the bottom of the soil furrow formed by the soil disturbed and by the previous rotary knives, and finally buried by the soil thrown up by the subsequent rotary tillage, and that the soil-preceding rice stubble would slowly break away from the rotary tillage trajectory to do curved motion, and would be buried by the airborne soil-preceding rice stubble due to the eccentricity of the center of mass of the soil-preceding rice stubble. Because the center of mass of the soil-preceding rice stubble in the air is biased toward the root end of the rotation of the two scenarios. By calculating the Spearman's rank correlation coefficients (0.91,0.84) between the simulation and the soil groove experiment trajectory, the accuracy of the contact motion of the simulation model was verified, and reliable contact parameters were provided for the discrete element simulation analysis of the preceding rice stubble rotary tillage trajectory regulation process.